The nucleotide sequence of the Paenibacillus curdlanolyticus B-6 xyn10A gene, encoding a xylanase Xyn10A, consists of 3,828 nucleotides encoding a protein of 1,276 amino acids with a predicted molecular mass of 142,726 Da. Sequence analysis indicated that Xyn10A is a multidomain enzyme comprising nine domains in the following order: three family 22 carbohydrate-binding modules (CBMs), a family 10 catalytic domain of glycosyl hydrolases (xylanase), a family 9 CBM, a glycine-rich region, and three surface layer homology (SLH) domains. Xyn10A was purified from a recombinant Escherichia coli by a single step of affinity purification on cellulose. It could effectively hydrolyze agricultural wastes and pure insoluble xylans, especially low substituted insoluble xylan. The hydrolysis products were a series of short-chain xylooligosaccharides, indicating that the purified enzyme was an endo-$eta$-1,4-xylanase. Xyn10A bound to various insoluble polysaccharides including Avicel, $alpha$-cellulose, insoluble birchwood and oat spelt xylans, chitin, and starches, and the cell wall fragments of P. curdlanolyticus B-6, indicating that both the CBM and the SLH domains are fully functioning in the Xyn10A. Removal of the CBMs from Xyn10A strongly reduced the ability of plant cell wall hydrolysis. These results suggested that the CBMs of Xyn10A play an important role in the hydrolysis of plant cell walls.